[Rivet] minutes of special LCG Generator Services monthly meeting - HepMC discussion and release planning

[Lynn Garren]

Thanks Andy.  Indeed there is a problem if you do not explicitly name 
any of the weights.  I am fixing that.

Tests are in place for the new range iterators, look in testHepMCIteration.

IteratorRange.h is deliberately simple. It was created to avoid a 
circular dependency problem, so we can't put the range classes there. 
How about if I rename GenEventIterators.h as just GenRange.h?

          Lynn


Andy Buckley wrote, On 05/11/2010 05:48 AM:
> On 06/04/10 00:00, Lynn Garren wrote:
>> The first beta of 2.06 is finally up.  I have only made the source code 
>> tarball available initially: 
>> http://lcgapp.cern.ch/project/simu/HepMC/download/
>>
>> In the unlikely event that there are no changes requested by any of you, 
>> I'll build the binaries.
> 
> Hi Lynn,
> 
> Apologies for taking a long time to report anything... it seems to work
> okay for me, but I've not had time to do any thorough testing. A few
> comments:
> 
> * Are there any tests/examples for the new features like labelled
> weights and range iterators? I couldn't find any, and they would be
> useful for testing. I didn't dig further than testHepMC.cc, though...
> maybe elsewhere?
> 
> * I had trouble reading in weights from a HepMC 2.05 ASCII file... is
> the HepMC 2.06 code able to handle the old, non-labelled weights? I
> think this is probably needed, at least until use of the naming feature
> is well-established.
> 
> * The new range iterators, for GenVertex as well as GenEvent are defined
> in GenEventIterators.h -- it took me some time to realise that the
> contents of this file aren't just limited to what the name indicates.
> Any chance they could go in the (already existing) IteratorRanges.h
> instead -- the name seems to fit, and it currently just contains one
> enum declaration, which itself is used by the new range functions.
> 
> * Do the weight labels actually propagate to the ASCII file? I tried
> generating some events and didn't see any corresponding entry in the
> file for empty labels, but maybe it isn't written out if the labels
> aren't set?
> 
> Cheers, and sorry again to take so long before moaning ;)
> Andy
>